Long-Term Performance of Divesting Firms and the Effect of Managerial Ownership

Introduction

Corporations that divested assets during the 1980s produced significant gains for their shareholders. The source of these gains has drawn the attention of researchers in recent years but remains an open empirical issue. Early studies mainly analyzed the wealth effects of divestitures and suggested, but did not specifically analyze, motives for asset sales.1 Recent cross-sectional studies of divestitures have suggested that gains arise from correcting poor investment decisions, such as improving corporate focus by removing assets unrelated to the firm's core operations or undoing failed acquisitions. In this study, we examine the long-term stock market performance and operating performance of divesting firms. We extend the cross-sectional analysis of divestitures by investigating whether gains reflect reduced agency costs and are therefore related to ownership and board structure.

Recent studies that analyze gains from divestitures include Kaplan and Weisbach (1992), who study the success of acquisitions and show that nearly half of the acquisitions in their sample were later divested, and about a third of these divested acquisitions were unsuccessful. They find positive returns at the announcement of the sell-off, but significantly higher returns for divestitures they classify as unsuccessful, suggesting that the divested assets impede the firms' other operations. John and Ofek (1995) see increasing corporate focus as the source of gain. They show that returns are higher, and that the firm experiences increased operating performance, when divestitures lead to a more focused firm. As an alternative to correcting overinvestment, Lang, Poulsen, and Stulz (1995) argue that managers value firm size and control, that divestitures are used to raise capital, and that gains come from reducing both information asymmetries and the costs of managerial discretion. They show that divestiture announcement returns depend on the use of the proceeds, and that positive returns occur when proceeds are paid out to creditors or shareholders. When proceeds are retained, announcement returns are insignificantly negative.

Even though studies show that shareholders gain from divestitures, and that the gains are consistent with increased efficiency, few studies have analyzed long-term operating performance following asset sales (John and Ofek 1995), and none have analyzed the long-term stock market performance, although Cusatis, Miles, and Woolridge (1993) and Desai and Jain (1999) study long-term stock market performance of parent firms following spinoffs. Examining long-term performance following asset sales is appropriate because managers could sell assets for reasons other than removing assets that do not fit with core operations and generate negative synergies. For example, Mitchell and Mulherin (1996) suggest that performance may not necessarily improve following takeovers and restructurings that result from industry shocks. Thus, post divestiture performance will provide further insight about the source of gains. Moreover, recent evidence suggests that the initial announcement may not fully reveal the motive for the sale. For example, Ikenberry et al. (1995) and Spiess and Afflick-Graves (1995) find evidence of underreaction to announcements of changes in the firms' financing policies, and Agrawal et al. (1992) find evidence of underreaction to announced changes in investment policies.

Our study finds that, on average, divesting firms underperform control sample firms over the two years preceding the divestiture measured by abnormal buy-and-hold returns, while outperforming the control firms over the three years following the divestiture. Divesting firms exhibit lower return on assets and market-to-book ratios before the divestiture and higher values of these performance measures following the divestiture. Announcement period returns are significantly positive when divesting firms exhibit negative abnormal buy-and-hold returns before the divestiture, and when post divestiture abnormal buy-and-hold returns are positive. Additional tests show a strong direct relation between long-term performance and managerial ownership in general, and CEO ownership in particular. Our evidence is consistent with the notion that gains arise from removing assets that generate negative synergies, and that the asset sale helps resolve agency problems.

Source of Gains and Ownership Structure

Theory suggests that asset sales create value for three reasons. First, asset sales improve the operational efficiency of the divesting firm by eliminating negative synergies. Second, for firms that suffer from severe costs associated with asymmetric information, debt overhang, or managerial discretion, asset sales are an efficient means of raising capital. Third, assets are sold to a firm that is willing to pay more for them than the assets are worth to the divesting firm. In the first two instances, higher values originate within the selling firm itself by correcting suboptimal investment policies that arise from agency problems when managers pursue their own objectives. In the third case, value originates in the buying firm and is transferred to the seller through the bargaining process and does not necessarily reflect resolution of problems within the selling firm.

Sources of Gains

An important source of gain in divestitures is the improvement in long-term performance that comes from eliminating negative synergies (Hite, Owers, and Rogers 1987 and John and Ofek 1995). Negative synergies arise, for example, when assets unrelated to core operations prevent a firm from focusing on its core competencies. Other examples include assets from a failed acquisition (Kaplan and Weisbach 1992), or operations that continually lose money under the current organizational structure and increasingly draw valuable resources from other parts of the firm. As asset sales remove these negative synergies, gains arise from improved efficiency in other parts of the firm. Thus, we expect that long-term performance improves after synergy-motivated divestitures.

A second endogenous source of gain comes when the divestiture reduces the costs related to market inefficiencies or agency costs. Lang, Poulsen, and Stulz (1995) suggest that managers value firm size and control, and that managers sell assets to raise capital when asymmetric information, debt overhang, or managerial discretion make an equity offering too expensive. Shareholders benefit because the announcement reduces information asymmetries and resolves uncertainty about the true value of the assets. Shareholders receive additional gains from reduced agency costs when the proceeds are used either to pay down debt, distributed as cash dividends, or to buy back stock. Unlike removing negative synergies, reducing the costs of asymmetric information produces a one-time gain, but not necessarily an improvement in long-term operating performance. When managers retain funds from asset sales and use them to pursue their own objectives, which is likely if managers value firm size and control, long-term performance might actually suffer. Thus, long-term performance will not necessarily improve, but could even deteriorate, when divestitures are used to raise capital.

Gains can also originate from the buyer. Hite, Owers, and Rogers (1987), among others, develop the efficient deployment argument, which suggests that divestitures move assets to higher-valued uses and shareholders capture some of the gains through effective bargaining. For example, the buyer creates gains from positive synergies by combining the divested assets with complementary resources or by managing the assets with a more efficient organizational structure. The efficient deployment argument suggests that the selling firm merely extracts part of the gains generated by the buyer.

The efficient deployment argument is silent about the nature of the divested assets and the use of the proceeds. As long as the buyer is willing to pay a premium for these assets, the divesting firm could be selling assets that have negative, zero, or positive net present value. Although the divesting firm is more likely to sell off poorly performing assets, Weston (1989) argues that this need not be the case. In addition, if managers of the divesting firm pursue their own objectives by using the proceeds in ways that harm shareholders, pay out the proceeds to shareholders, or reinvest the proceeds in positive net present value projects, the divestiture will likely lead to negative, zero, or positive changes in long term performance. Thus, the efficient deployment argument offers no clear-cut prediction about improvements in the selling firm's long-term performance.2

Ownership Structure and Gains

Theory suggests that gains from divesting assets arise from resolving agency problems that exist when internal controls are weak (Jensen 1993). Gains, therefore, will be related to managerial stock ownership, but the relation is not straightforward. Gains and ownership will be directly related if divestiture gains reflect expected improvements in operations and how the proceeds from the sale will be used. High levels of stock ownership will provide managers the incentive to improve operations previously affected by negative synergies and to invest the proceeds in ways that increase shareholder value. Moreover, managerial ownership will provide incentives to avoid the free cash flow problems described by Jensen (1986), perhaps by using the proceeds to pay down debt or repurchase stock.

An alternative view suggests that divestiture gains and managerial ownership will be inversely related if the gains reflect the magnitude of the negative synergies generated by the divested assets. Managers with significant ownership have the incentive to divest before the negative synergies become too large. Managers who own too few shares, however, have little economic incentive to sell assets that generate negative synergies, and may have incentives to retain these assets if they justify a higher level of compensation or reduce the manager's risk. Moreover, as Boot (1992) suggests, these managers could be reluctant to sell because the sale would highlight past mistakes or affect reputation. Managers will agree to sell, perhaps with the board's encouragement, when gains from the sale become too large to ignore, thus creating a negative relation between gains and ownership.

A third view suggests that managerial ownership could be unrelated to gains from divestitures when external factors such as economic shocks motivate the sale (Mitchell and Mulherin 1996). Here, the divestiture does not resolve agency problems that are usually associated with overinvestment when managers pursue their own objectives. Thus, even though an industry-shock motivated divestiture produces a gain, there is no necessary reason to believe that it would be related to managerial ownership.

Data and Methods

Sample

We collect an initial sample of 951 divestitures from Mergers and Acquisitions occurring from 1981 through 1995. The sample was reduced to 588 by requiring that the divestiture be reported in the Wall Street Journal and have stock returns on the Center for Research on security Prices (CRSP) files.3 We further required that information about board structure and managerial stock ownership for the year of the divestiture be available from Q-file proxy statements or the sec's Edgar database, which reduced the sample to 545. Using Compustat, we identified control firms matched by size and market-to-book ratio (both within 80 percent to 120 percent of the divesting firm's value), further reducing the sample to 300. The availability of proxy data, financial data, and long-term returns for the control firms reduced the sample to its final size of 213 divestitures.

Following the scheme used by Baysinger and Butler (1985), Byrd and Hickman (1992), and Brickley, Coles, and Terry (1994), among others, we classify the selling firm's directors based on their economic or personal ties to the firm. Directors are classified as inside if they are current or past employees or are related to the firm's chief executive or part of the founding family. Directors fall into the gray area if they have a close but indirect relation with the firm such as consultants, lawyers, or executives of firms with a business relationship with the firm. Otherwise, directors are deemed independent outside or unaffiliated directors such as executives of unrelated firms, private investors, or directors from outside the business community.

We calculate announcement period excess returns using the market model with the CRSP value-weighted index and estimate parameters over a 240-day period ending 60 days before the divestiture was reported in the Wall Street Journal. We report excess returns for the two days surrounding the Wall Street Journal announcement. In our tests for statistical significance, we calculate z-statistics using the technique described by Dodd and Warner (1983). We calculate long-term returns following the technique described in the next section.

Long-Term Performance

For long-term performance we follow Barber and Lyon (1997) and compare holding period returns of our sample firms with holding period returns from a control group matched by size and book-to-market ratio. We match on size and book-to-market, rather than SIC defined industries, because studies by Fama and French (1992), and others, have shown that these variables explain cross-sectional returns. Industry matching using SIC codes is problematic because contemporaneous SIC codes are generally not available, a firm's SIC codes may change over time, and the classification depends on the source (Kahle and Walkling 1996). Moreover, Barber and Lyon (1997) find that abnormal long-term returns calculated using a market index are biased, while a size and book-to-market matched control firm benchmark provides well-specified test statistics. They also suggest that buy-and-hold abnormal returns are superior to cumulative abnormal returns.4 Accordingly, we find control firms to match our divesting firms by first finding all firms with market value of equity within 80 percent to 120 percent of the sample firm's value, and then selecting the firm with the closest book-to-market ratio.

Empirical Results

Sample Description

Table 1 presents the distribution of asset sales by year and median values of the dollar value of the transaction and the value of the transaction relative to the market equity value of the parent firm. Table 2 presents descriptive statistics for the sample of divestitures. Divested assets were valued at $190 million (median $100 million). These values are nearly four times as large as those reported by Hite and Vetsuypens (1989), about twice the value reported by Lang, Poulsen, and Stulz (1995), but only half the value reported by John and Ofek (1995). The selling firm's equity was valued at $7.8 billion (median $2.4 billion), compared to $4.6 billion reported by John and Ofek (1995) and $0.9 billion reported by Lang et al. (1995). Different sample selection procedures and our more recent sample account for the size differences. On average, the value of the transaction is 26.8 percent of the firm's equity before the sale (median 5.4 percent). In contrast, John and Ofek (1995) report a relative value of 39.4 percent (median 15.3 percent), while Lang et al. report a relative value of 69 percent (median 23 percent). Lang et al. also show that the median relative value for firms that retain the proceeds of the asset sale is 13 percent, which is significantly smaller than the 42 percent relative value for firms that pay out the proceeds. This suggests that when the relative value is low, such as in our sample, firms are more likely to retain the proceeds (as do 70 percent of our firms) and increase the costs of managerial discretion.

Table 2 also shows various corporate governance metrics. CEOs of the divesting firm owned on average 3.27 percent of the firm's stock (median 0.27 percent), while officers and directors as a group owned 7.07 percent (median 1.80 percent). Inside directors, including the CEO, owned 5.80 percent (median 0.81 percent), while unaffiliated outside directors owned only 0.79 percent (median 0.04 percent) of the firm's shares. By contrast, Lang et al. (1995) report managerial ownership of 13 percent (median 8 percent) for their sample of much smaller firms that engaged in asset sales. For a random sample of firms, Denis and Sarin (1999) report mean ownership by officers and directors of 15.74 percent (median 8.08 percent) and by the CEO of 7.22 percent (median 0.30 percent). Their sample, however, consists of much smaller firms with mean equity value of $435 million (median $59 million). Thus, the low ownership levels in our sample probably reflect the larger firm size. In our sample, although outside directors own few shares, board structure shows that outside directors on average constitute 52.7 percent (median 53.8 percent) of board members, and make up more than half of the board in 54.0 percent of sample firms. Nearly half of the firms (49.8 percent) have outside 5 percent blockholders. Finally, as with previous studies, divestiture announcements produce significantly positive two-day (-1, 0) excess returns (0.429 percent, z=2.00).

As a test of the Lang, Poulson, and Stulz (1995) financing hypothesis, we searched annual reports and 10K reports to find specific mention of the use of the proceeds from the sale. Their model suggests that firms that pay out the proceeds should have higher announcement period returns and higher levels of managerial ownership than firms that retain the proceeds. We found that 61 firms specifically mentioned that the proceeds would be used to pay down debt or repurchase shares. The remaining 152 firms made no mention of how the proceeds would be used or indicated that the proceeds would be used to acquire assets. Among the payout sample, two-day announcement period returns were 0.755 percent (z = 1.84) and ownership by officers and directors was 8.92 percent, while for the retained sample returns were 0.298 percent (z = 1.20) and ownership was 6.32 percent. The differences, however, are not significant (t = 0.66 and t = 1.07, respectively). Thus, in our sample, how the firm uses the proceeds from the divestiture has no significant explanatory power.

Because the distribution of relative values is highly skewed, we partitioned the sample by the median relative value (5.4 percent). The last two columns in Table 2 show that, not surprisingly, the above median group is characterized by smaller firms (average equity value $1.5 billion) selling substantial assets (average $237 million), while the below median group represents larger firms (average equity $14 billion) selling minor assets (average value $143 million). Moreover, consistent with Demsetz and Lehn (1985), the smaller firms in the above median group have significantly higher managerial stock ownership than the larger firms in the below median group. CEOs in the above median group, for example, control 6 percent of the firms' shares versus 0.55 percent ownership for the above median group. Officers and directors in the above median firms have more than three times the ownership (11 percent versus 3 percent) than their counterparts in the below median firms. Finally, announcement period abnormal returns are about three times larger (0.662 percent versus 0.199 percent) for the above median firms, but the difference is not statistically significant.

Long-Term Performance

Table 3, Panel A, reports buy and hold abnormal (BHAR) returns for various intervals from two years before to three years after the divestiture. We calculate abnormal returns following Barber and Lyon (1997) by subtracting holding period returns for a size and market-to-book ratio matched control firm from the divesting firm's holding period returns. During the two years preceding the divestiture, divesting firms significantly underperformed control sample firms by 12.6 percent (t = 2.33). Interestingly, the divestiture seems to halt the underperformance. Divesting firms' performances match control sample firms' performance for the first two years following the divestiture, and slightly outperform by 9.6 percent (t = 1.25) over the three years following the divestiture. The statistically neutral returns following the sale contrast with Cusatis et al. (1993), who find strong positive matched-firm adjusted returns to parent firms following spinoffs, which they attribute to post-spinoff takeover activity, and Desai and Jain (1999), who find positive buy and hold returns for firms that increase focus following the spinoff. Although the 12.6 percent underperformance seems economically significant, the statistical reliability of the estimate could be subject to some skepticism. Mitchell and Stafford (2000) argue that the statistical reliability of BHARs is overstated because estimated standard deviations used to form test statistics ignore cross-sectional correlations among BHAR that arise from, for example, time and industry clustering of events. Nevertheless, the pattern of buy-and-hold abnormal returns shown in panel A provides support for the notion that performance improves after the divestiture because the sale removes negative synergies. Insignificant returns after the sale, however, are consistent with the Mitchell and Mulherin (1996) argument that performance will not necessarily improve after restructuring motivated by industry shocks. We investigate this further in panels B and C.5

In panel B, we report changes in the market-to-book ratio and return on assets (EBITD/TA) over the same intervals that we report BHAR in panel A. For operating cash flow we use earnings before interest, depreciation, and amortization (Compustat data item 13). We report excess values of the ratios, which are the change in the divesting firm's ratio less the change in the matched control firm's ratio. In the two years leading up to the divestiture, the change in the divesting firm's market-to-book ratio was not significantly different from the change in the matched control firm's ratio. But following the divestiture, the divesting firm's market-to-book ratio increased significantly more than the control firm's market-to-book ratio. The difference is significant over each period, but at a declining level of significance. This pattern is consistent with the notion that the divestiture precipitates a change in performance, and hence, valuation of the firm.6 The actual ratios are shown in panel C. We see that from year 0 to year 3, for example, the divesting firm's market-to-book ratio increased from 2.3 to 4.5, while the matched control firm's ratio increased from 2.3 to 3.4. At year 3, pairwise t-tests (t = 1.86) and Wilcoxon signed-ranks tests (z = 2.65) show that the difference in the ratios is significant.

Also shown in panel B are excess changes in return on assets (EBITD/TA). Changes in EBITD/TA show that the divestiture had much more effect on the firm's performance than suggested by BHAR or the market-to-book ratio. Over the two years before the divestiture, divesting firms experienced significantly larger reductions in return on assets (-0.011, t = -2.20) compared to control firms. After the divestiture, the reverse occurred, and divesting firms experienced significantly larger increases in return on assets. For example, from year 0 to 3 the increase in return on assets for divesting firms significantly outpaced control firms by 1.8 percent (t = 3.00). Panel C reports the actual values for each year. For divesting firms, return on assets is consistently lower than for the matched control firms over the two years before the divestiture, and consistently higher after the divestiture. By year 3, divesting firms have a return on assets of 15.4 percent, which is significantly higher than the 13.8 percent experienced by the matched control firms.

The pattern of BHAR and increases in market-to-book ratios and return on assets following divestitures that we document in Table 3 suggest that the divestiture allowed the firm to change from a below-average performer to an above-average performer. This is consistent with the view that gains arise from removing assets that cause negative synergies. Moreover, changes in performance surrounding divestitures suggests that the sales were not solely motivated by the prospect of selling assets to a buyer willing to pay a high price, nor solely a result of industry shocks. Our results contrast with Powers (2002), who finds that in equity carve-outs parent firms carve out high-growth, high-value divisions. In the next section, we investigate whether the change in performance was related to the resolution of agency problems.

Announcement Returns and Long-Term Performance

In Table 4 we provide preliminary evidence about whether the poor performance leading up to the divestiture is an indication of agency problems, and whether the improved performance after the divestiture reflects the resolution of these problems. In the table we divide the sample into four groups. In panel A we first divide the sample by the sign of buy-and-hold abnormal returns over the two years preceding the divestiture. If poor performance over this period reflects agency problems within the firm, then we would expect that managerial ownership would be lower for firms experiencing negative buy-and-hold abnormal returns. Panel A shows that this is not the case; there is no significant difference (t = 0.83) in stock ownership by officers and directors between firms that have positive BHAR (ownership = 7.97 percent) and firms that experience negative BHAR (ownership = 6.27 percent). Interestingly, we see that announcement period excess returns are significant (0.755 percent, z = 2.45) only when buy-and-hold abnormal returns are negative, and firms in this group are about half the size as the positive returns group. The results in panel A suggest that gains arise from removing negative synergies, but the negative synergies may not be associated with low managerial ownership.7

In panel B, we divide the sample by the sign of buy-and-hold abnormal returns over the two years following the divestiture. As in panel A, managerial ownership does not differ significantly (t = 1.22) between groups, although the positive BHAR ownership level is higher by nearly 2.5 percent. Negative buy-and-hold abnormal returns are still associated with smaller firms, although the relative size of the transaction is larger. Consistent with market efficiency, positive buy-and-hold abnormal returns are associated with significantly positive announcement period returns (0.842 percent, z = 2.36).

Although the results in Table 4 suggest that long-term performance seems unrelated to managerial ownership, and therefore agency problems, we should be cautious about drawing strong inferences. One concern is that our ownership levels are measured at the end of, and not the start of, a two-year period characterized by poor performance. In addition, ownership levels could change following the divestiture. Denis and Sarin (1999), for example, find that managerial ownership is negatively related to market-adjusted stock returns. Thus, our levels could be overstated. They also find that changes in ownership typically occur along with or following restructuring activities. Notwithstanding these concerns, our univariate estimates presented in Table 4 could mask some important cross-sectional differences. We examine this next with our regression analysis.

Regression Analysis

Theory suggests that higher levels of stock ownership by managers and monitoring by the board of directors provide managers with more incentive to avoid investment strategies that benefit them more than shareholders. Examples include assets unrelated to the firms' core business, assets that are part of failed acquisitions, or manager-specific assets. We investigate this issue by using cross-sectional regressions to examine the effect that managerial ownership and board structure has on the divesting firm's long-term market performance. If improvements in long-term performance are related to resolving agency problems, then managerial ownership and board structure should have significant explanatory power.

Table 5 shows estimates when we regress abnormal buy and hold returns over the two years following the divestiture on various measure of ownership and board structure.8 In regression 1, we isolate the effect of managerial ownership while controlling for firm size, transaction value, and use of proceeds. Ownership by officers and directors has a significant (t = 3.33) positive effect on long-term returns, consistent with the view that divestitures, in part, resolve agency problems. The value of the transaction has a significant effect on long-term returns, suggesting that the sale reduces overinvestment. Whether the firm pays out the proceeds (use of proceeds dummy equals 1) to pay down debt or repurchase shares, or retains the proceeds, has no impact on long-term returns. In regression 2, we gain additional insight into the role of managerial ownership by decomposing ownership by officers and directors into ownership by the CEO and each classification of director. Regression estimates show that ownership by the CEO is the only significant (t = 2.68) ownership variable, consistent with the notion that stock ownership provides the CEO a strong economic incentive to improve operations. Ownership by unaffiliated outside directors is positive but insignificant (t = 1.30), providing weak support to the notion that monitoring by these directors enhances firm performance.

We further explore the influence of monitoring by the board of directors and outside 5 percent blockholders in regressions 3 and 4. If managers own too few shares to align their interests with shareholders, then monitoring by the board, especially unaffiliated outside directors, and blockholders should help resolve agency problems. We use dummy variables to identify where unaffiliated outside directors constitute more than half the board and firms that have outside (without board representation) 5 percent blockholders. In regression 3, ownership by officers and directors remains strongly significant (t = 3.46), and the presence of an outside blockholder has a marginally significant (t = 1.56) positive effect on long-term returns. The outside director dummy variable, while positive, is not significant. The influence of outside blockholders weakens when, in regression 4, we use separate measures of ownership by the CEO and directors by affiliation. In regression 4, CEO ownership is the only significant (t = 2.78) ownership variable. These estimates suggest that, after controlling for monitoring by outside blockholders and unaffiliated directors, stock ownership by the CEO has the most direct influence on long-term performance.

To compare to earlier studies that focus on announcement period returns, e.g., Lang et al. (1995), and that usually find no significant relation between returns and managerial ownership, we estimated a regression similar to regression 3 but using two-day announcement returns as the dependent variable (results are not reported in the tables). The ownership coefficient in this regression was positive but not significant (t = 1.42). The absence of a significant relation is consistent with the view that divestitures are a response to economic shocks and unrelated to managerial ownership. It would also be consistent with the view that the positive and negative relations between ownership and gains hypothesized earlier offset one another.

In regression 5 we revisit ownership by officers and directors, but now include the interactive term use of proceeds dummy x ownership to see whether ownership is related to long-term returns when managers choose to retain the proceeds and possibly exacerbate agency problems. The ownership coefficient is weaker but still significant (t = 1.81 versus t = 3.33 in regression 1) when proceeds are retained. When the ownership and interactive term coefficients are combined (0.019 = 0.008+0.011), the combined coefficient is highly significant (t = 3.49).9 Thus, ownership is weakly important in explaining long-term returns when proceeds are retained, and much more important when proceeds are paid out. This result is consistent with the view that payout of the proceeds is more likely when managerial ownership is high and agency costs are low.

Table 2 shows that when the sample is partitioned at the median relative value of the transaction, managerial ownership is significantly higher for the above median group. To examine the power of managerial ownership to explain abnormal buy and hold returns, we estimated regressions 1 and 5 separately for the above-median and below-median groups. Results, not reported in the tables, show that in the below-median relative value group, ownership had no explanatory power. However, in the above-median relative value group, similar to regressions 1 and 5 reported in Table 5, managerial ownership by officers and directors remains significant in regression 1 (p = 0.00) and regression 5 (p = 0.06). These results further support the notion that managerial ownership is an important factor in asset sales especially when the sale represents a substantial fraction of the firm's value.

Our regression results suggest that, for our sample, improvement in long-term performance as measured by abnormal buy-and-hold returns over the two years following the divestiture is significantly influenced by CEO stock ownership. We find no evidence consistent with the view that divestiture gains and managerial ownership should be negatively related. Our estimates are consistent with the notion that resolving agency problems plays a part in the gains from divestitures.

Conclusions

Divestitures can resolve agency problems by reducing overinvestment, exacerbate agency problems by providing managers cash to pursue their own objectives, or even be unrelated to agency problems if the sale is in response to economic shocks. Well-documented small positive announcement period returns suggest that divestitures generate gains for shareholders, but the source of these gains still remains an open issue. In this study, we investigate the long-term performance of firms that divest assets in order to produce further insights about the source of these gains, and whether resolving agency problems explains some of the gains. Firms that sell assets that cause negative synergies should experience improved performance, but firms that divest assets to raise capital or in response to economic shocks may not see improved performance. In this study, we find that divesting firms underperform control sample firms during the two years preceding the divestiture whether performance is measured using buy-and-hold abnormal returns, market-to-book ratios, or return on assets. In the three years following the divestiture, divesting firms outperform control sample firms. These results support the notion that divestitures improve the firm's operations, perhaps by removing negative synergies. Further analysis suggests that the poor performance preceding the divestiture is unrelated to managerial ownership, but that the post-divestiture performance is strongly related to stock ownership by the CEO. Overall, our results lend support to the argument that divestitures eliminate negative synergies, that these negative synergies are not solely the consequence of agency problems, but that managerial ownership provides strong incentives to improve operations following the divestiture.